Search

Built-in tests (BITs) are widely used in manufacturing and production systems to find whether system failures occur, whereas the problem of BIT false alarms caused by intermittent failures adds to much trouble for the precise failure detection and diagnosis. Fighting with false alarms caused by intermittent failures is an urgent issue. However, the nature and temporal regularity of intermittent failures are not fully exploited, as well as the relationship between intermittent failure and BIT false alarms. The present paper introduces the method of constructing failure test profile for false alarm assessments. Probabilistic models are proposed of the failure evolution process, as well as the interactions between intermittent failures and false alarms. The false alarm time expectation is derived with the given model, serving as the foundation for the optimisation problem to find the best test threshold to enable the highest BIT capability. A numerical analysis is made to illustrate the proposed model and examine the threshold determination method. An application study is also carried out to show how the model can be applicable in real engineering practices. [ABSTRACT FROM PUBLISHER]

State-of-health (SOH) assessment and remaining useful life (RUL) estimation are among the key issues in prognostics and health management (PHM) for electronic systems. Unlike mechanical systems, the homogeneity of the fault modes is quite low for electronic systems. Seen at the system level, there are also multiple and uncertain fault modes for electronic products. In this paper, we propose a novel methodology for system-level SOH assessment and RUL estimation inspired by the quantum mechanics disciplines, where there is no requirement to distinguish the fault modes from the fault development patterns. It is developed on the analogy that the healthy data points tend to move to the lower potential energy positions, and the subhealthy or unhealthy ones tend to be repelled from such positions. The fault development paths are located on the potential surface. The Fermi–Dirac health descriptor (FDHD) is defined based on the pseudowave function and potential function. Based on FDHD, the SOH assessment and RUL estimation algorithm are developed to track the existing fault development paths. Finally, the proposed method is verified in an experiment using power conversion board with a detailed analysis of the SOH assessment results and RUL estimation performances. [ABSTRACT FROM AUTHOR]

Fault diagnosis of analog circuits is more challenging compared with digital circuits as a result of the parametric deviation and the difficulty in signal discretization. There still lacks effective approaches to realize reliable fault detection and isolation for a comprehensive diagnosis. A new fault diagnosis technique called multi-valued Fisher's fuzzy decision tree (MFFDT) is proposed in this paper to solve the problem. This technique uses the decision tree as the diagnosis model and incorporates the Fisher's linear discriminant principles. The fuzzification mechanism is devised to discretize the input monitoring data. The proposed MFFDT method is composed of two aspects: decision tree training and real fault diagnosis processes. The former uses the benchmark data to train a decision tree, while the latter sends the monitoring data into the decision tree to generate diagnosis results. The proposed method is validated using simulated data and the real-time data for an active filter circuit and an audio amplifying circuit. The comparative analysis is also presented to evaluate diagnosis performances. Copyright © 2015 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Built-in tests (BITs) are widely used in mechanical systems to detect and diagnose a fault, whereas the BIT false alarms bring much trouble for precise fault diagnosis and logistics/maintenance arrangement. The false alarm phenomenon is related to the degradation over time, and the false alarm evolution process can be typically divided into three stages. This paper proposes a condition-based multistage false alarm detection and reduction method for mechanical systems. The stages are clarified according to the degradation level and the false alarm severity. The dividing boundaries of the stages are optimized using soft margin one-versus-rest support vector machine (SVM) classifiers. The associated intermediate stage is the intense period of false alarms, and the dynamic Bayesian network inference model is developed to satisfy the requirements of accurate false alarm diagnosis. To achieve the goal of false alarm suppression, the top-level BIT outputs are updated with the original BIT alarms and the identified probable states. Finally, the proposed approach is demonstrated in the application study of a milling machine and the well-round experimental results are analyzed. [ABSTRACT FROM PUBLISHER]